Thermal Model Performance for the James Webb Space ...
Transcript of Thermal Model Performance for the James Webb Space ...
48th International Conference on Environmental Systems ICES-2018-35 8-12 July 2018, Albuquerque, New Mexico
Copyright © 2018 NASA Goddard Space Flight Center
Thermal Model Performance for the James Webb Space
Telescope OTIS Cryo-Vacuum Test
Kan Yang1, Stuart D. Glazer2, Shaun R. Thomson3, and Lee D. Feinberg4
NASA Goddard Space Flight Center, Greenbelt MD 20771
William Burt5, Brian J. Comber6, and Wes Ousley7
Genesis Engineering Solutions Inc., Lanham, MD 20706
Randy Franck8
Ball Aerospace, Boulder, CO 80301
The James Webb Space Telescope (JWST), set to launch in mid-2020, is currently undergoing
a series of system-level environmental tests to verify its workmanship and end-to-end
functionality. As part of this series, the Optical Telescope Element and Integrated Science
Instrument Module (OTIS) Cryo-Vacuum (CV) test, the most complex cryogenic test executed
to date by NASA, has recently been completed at the Johnson Space Center’s Chamber A
facility. The OTIS CV test was intended as a comprehensive test of the integrated instrument
and telescope systems to fully understand its optical, structural, and thermal performance
within its intended flight environment. Due to its complexity, extensive pre-test planning was
required to ensure payload safety and compliance with all limits and constraints. A system-
level pre-test thermal model was constructed which fully captured the behavior of the payload,
ground support equipment, and surrounding test chamber. This thermal model simulated
both the transient cooldown to and warmup from a 20 K flight-like environment, as well as
predicted the payload performance at cryo-stable conditions. The current work is an
assessment of thermal model pre-test prediction performance against actual payload response
during the OTIS CV test. Overall, the thermal model performed exceedingly well at predicting
schedule and payload response. Looking in depth, this work examines both the benefits and
shortcomings of assumptions made pre-test to simplify model execution when compared
against test data. It explores in detail the role of temperature-dependent emissivities during
transition to cryogenic temperatures, as well as the impact that model geometry
simplifications have on tracking of critical hardware limits and constraints. This work
concludes with a list of recommendations to improve the accuracy of thermal modeling for
future large cryogenic tests. The insight gained from the OTIS CV test thermal modeling will
benefit planning and execution for upcoming cryogenic missions.
Nomenclature
AOS = Aft Optical System
ACF = Auto-Collimating Flat
ADIR = Aft Deployable ISIM Radiator
ASPA = Aft optical system Source Plate Assembly
BP = Back Plane
1 Thermal Engineer, JWST Project, NASA/GSFC, Code 545 2 Lead Thermal Systems Engineer, JWST ISIM Element, OTIS Thermal Lead Engineer, NASA/GSFC, Code 545 3 Lead Thermal Systems Engineer, JWST Project, NASA/GSFC, Code 545 4 Optical Telescope Element Manager, JWST Project, NASA/GSFC, Code 550 5 Lead Cryogenic Thermal Systems Engineer, 4501 Boston Way, Lanham, MD 20706 6 Thermal Engineer, 4501 Boston Way, Lanham, MD 20706 7 Thermal Systems Engineer, 4501 Boston Way, Lanham, MD 20706 8 Lead Thermal Systems Engineer, 1600 Commerce St, Boulder, CO 80301
International Conference on Environmental Systems
2
BSF = Backplane Support Fixture
CoCOA = Center of Curvature Optical Assembly
CPP = Cryo-Pumping Panels, separate cold panels between the helium and LN2 shrouds at NASA JSC
CSA = Canadian Space Agency
CTE = Coefficient of thermal expansion
CV = Cryo-Vacuum
ΔT = Change in temperature
DTA = Deployable Tower Assembly
DSERS = Deep Space Environment Radiative Sink
EC = European Consortium
ESA = European Space Agency
FGS = Fine Guidance Sensor
FIR = Fixed ISIM Radiator
FPA = Focal Plane Arrays
FSM = Fine Steering Mirror
GSE = Ground Support Equipment
GSFC = NASA Goddard Space Flight Center
HOSS = Hardpoint and Offload Support Structure
IEC = ISIM Electronics Compartment
IR = Infrared
ISIM = Integrated Science Instrument Module, which contains the Science Instruments (SIs)
JSC = NASA Johnson Space Center
JWST = James Webb Space Telescope
K = Kelvin
L&Cs = Limits and Constraints
L5 = Layer 5 Sunshield simulator
LN2 = Liquid Nitrogen
LRM = Launch Release Mechanism
MIRI = Mid-Infrared Instrument
MLI = Multi-Layer Insulation
NASA = National Aeronautics and Space Administration
NGAS = Northrop Grumman Aerospace Systems
NIRCam = Near-Infrared Camera instrument
NIRSpec = Near-Infrared Spectrograph instrument
OA = Optical Assembly
OGSE = Optical Ground Support Equipment, a series of pre-OTIS optical pathfinder tests
OTE = Optical Telescope Element
OTIS = Optical Telescope Element plus Integrated Science Instrument Module (OTE + ISIM)
PG = PhotoGrammetry cameras
POM = Instrument Pick-Off Mirror
PM = Primary Mirror(s)
PMSA = Primary Mirror Segment Assembly
PMBSS = Primary Mirror Backplane Support Structure (BSF + BP)
SI = Science Instrument
SINDA = Systems Improved Numerical Differential Analyzer modeling tool
SM = Secondary Mirror
SMA = Secondary Mirror Assembly
SMSS = Secondary Mirror Support Structure
SVTS = Space Vehicle Thermal Simulator
TM = Tertiary Mirror
TPF = Thermal Pathfinder test
W = Watt(s)
International Conference on Environmental Systems
3
I. Introduction
HE James Webb Space Telescope (JWST) is NASA’s next-generation space telescope. As a collaboration
between NASA, the European Space Agency (ESA), the European Consortium (EC), the Canadian Space Agency
(CSA), and partners in industry and academia, JWST (Figure 1) will provide scientists with unprecedented capabilities
in the mid-to-near infrared bands to image the formation of the first stars and galaxies in the universe and to capture
exoplanets. After launch in 2020 atop an Ariane 5 launch vehicle from Kourou, French Guiana, JWST will undergo a
series of deployments before entering orbit at the Earth-Sun L2 Lagrange point, about 1.5 million km from Earth. A
tennis-court-sized, five layer sunshield will protect the optical telescope and instruments from solar radiation, allowing
the cold “telescope” portion of the spacecraft to cool to cryogenic temperatures for its science operations.
JWST has been undergoing extensive cryogenic vacuum tests to verify its ability to withstand the harsh space
environment. The size of the fully-deployed observatory makes it prohibitive to test in any existing facility. To ensure
that the telescope functions as planned and achieves its design cryogenic temperatures in the space environment,
optical and cryo-vacuum verification via two separate system-level assemblies is necessary before its final integration.
These tests include the spacecraft bus and the sunshield, comprising the spacecraft element (SCE); and the science
instruments and optics, which together form OTIS. As these separate system-level tests replace the full observatory-
level test, observatory-level verification is
achieved with thermal models correlated through
system-level testing. However, this requires an
increased level of rigor in the system-level testing
to ensure that models can accurately predict the
behavior seen at the subsystem and system levels,
since JWST relies on analysis to confirm that the
observatory-level assembly works as intended.
The OTIS cryo-vacuum (CV) test at NASA
JSC’s Chamber A facility took place from July
through October of 2017, and allowed for the
system-level verification of both the Optical
Telescope Element (OTE) and the Integrated
Science Instrument Module (ISIM), essentially
merging the telescope and its four instruments
into a fully-integrated science package1,2. OTIS
CV was the culmination of years of subsystem-
level OTE and ISIM cryo-vacuum tests: two
Optical Ground Support Equipment (OGSE)3 tests
and one Thermal Pathfinder (TPF)4-6 risk-reduction test at NASA’s Johnson Space Center (JSC) for OTE; and three
successive ISIM cryo-vacuum tests7 at NASA’s Goddard Space Flight Center (GSFC). OTIS allowed for end-to-end
optical testing in the appropriate thermal environment for the combined assembly of the optical telescope and
instruments, as well as validation of the thermal design and its ability to achieve expected cryogenic balance
temperatures. It further served as a verification of workmanship and as a crosscheck of the assembled and integrated
models. For this CV test, there were five thermal control objectives:
1. To preserve hardware integrity upon transition to cryogenic thermal balance (cryo-balance) conditions and
transition back to ambient temperatures by respecting all imposed limits and constraints;
2. To achieve the simulated on-orbit payload temperature levels and stability for optical, mechanical, and
instrument tests;
3. To predict and measure thermal balance test data for model crosscheck, both on ISIM and OTE components;
4. To achieve a workmanship thermal conductance assessment of the flight instrument heat straps which for the
first time would be connecting all the payload flight instruments and radiators;
5. To achieve test timeline optimization by executing the OTIS CV cooldown and warmup in a time-efficient
manner.
These thermal control objectives ensured that the test could be executed in a safe, efficient manner over its 100-
day duration. However, meeting these objectives required significant predictive capability to address all of the imposed
T
Figure 1. The James Webb Space Telescope and its major
system-level assemblies
International Conference on Environmental Systems
4
Limits and Constraints (L&Cs), including structural gradient, temperature rate-of-change, contamination control, and
survival limits. To this end, the OTIS CV thermal model was developed, consisting of an integrated system-level
model of the OTIS payload from Northrop Grumman Aerospace Systems (NGAS) with substantial inputs from NASA
GSFC, detailed optical component models from Ball Aerospace, and a reduced Ground Support Equipment (GSE)
model, including the chamber and facilities, provided by Harris Corporation. The OTIS CV thermal model was
generated with Thermal Desktop and SINDA/FLUINT version 5.6, contains over 84,000 nodes, and typically requires
more than a week of wall-clock time to run a simulated cooldown profile of 35 days or warmup profile of 20 days.
The OTIS CV thermal model predictions were used to develop appropriate cooldown and warmup procedures
employing the appropriate thermal control surfaces and methods to transition the OTIS payload from ambient to its
cryogenic test conditions and back to ambient8. The model was also used to evaluate critical L&Cs: those that lacked
flight sensors to monitor real-time during the OTIS CV test were predicted pre-test with sufficient margin to ensure
that they could be met during transitions.
The current work will discuss the OTIS CV thermal model’s performance against as-measured test data, especially
with regards to cooldown and warmup behavior of the payload itself. The system-level model made generally accurate
predictions for test duration and hardware performance against L&Cs. However significant simplifications and
assumptions needed to be made to ensure timely model predictions and reduce runtime. The lower model complexity
and resolution in certain areas also resulted in deviations of model predictions from actual hardware behavior. A
discussion of test configuration is provided in the next section to describe the OTIS CV test before examining the
most impactful decisions made with regards to thermal modeling.
II. Test Configuration
The OTIS payload encompasses all components of the observatory that lie above the plane of the coldest sunshield
layer. Figure 2 shows the OTIS payload in its test configuration at JSC; note the chamber coordinate system in so-
called V-coordinates. In the optical path for the telescope element, light reflects off of the 18 gold-coated beryllium
Primary Mirror Segments comprising the Primary Mirror Segment Assembly (PMSA), then the Secondary Mirror
Assembly (SMA), and lastly enters the Aft Optics Subsystem (AOS), where it travels from the Tertiary Mirror (TM)
to the Fine Steering Mirror (FSM) to the pick-off mirrors for each separate science instrument within the ISIM
enclosure. The ISIM consists of four instruments mounted to the rigid composite ISIM structure: the Near-Infrared
Camera (NIRCam), developed by NASA and the University of Arizona; the Near-Infrared Spectrograph (NIRSpec),
a collaboration between the ESA and NASA, separated into its Optical Assembly (OA) and Focal Plane Assembly
(FPA) components; the Fine Guidance Sensor (FGS), provided by the Canadian Space Agency; and the Mid-Infrared
Instrument (MIRI), jointly sponsored by ESA and the European Consortium.
Structurally, the PMSAs and the AOS are supported by the Backplane (BP), a large thermal insulation-wrapped
M55J composite structure comprised of a center section and foldable wings. ISIM is kinematically mounted to the
Backplane Support Fixture (BSF), a truss-like composite structure. Together, BP and BSF form the Primary Mirror
Backplane Support Structure (PMBSS), the mechanical framework which holds the telescope element and instruments
in place. The entire telescope assembly is then interfaced with the flight spacecraft bus and sunshield via the
Deployable Tower Assembly (DTA). On the –V1 side of the payload, the room-temperature ISIM Electronics
Compartment (IEC) controls and powers the instruments via each of their electronics boxes, and a set of harnesses
with their own Harness Radiator connects the IEC to the ISIM. Mounted on the –V1/+V3 side of the spacecraft, the
Fixed ISIM Radiator (FIR)9 and Aft Deployable ISIM Radiator (ADIR) passively reject heat from the instruments.
The JSC Chamber A thermal-vacuum test facility, the second largest thermal vacuum facility in the United States,
is a 16.8m diameter x 27.4m high cylindrical vacuum vessel. It was modified extensively to accommodate the OTIS
payload and to simulate a flight-like thermal environment10. Figure 3(a) shows the cross-sectioned CAD model of the
OTIS test payload cocooned within a series of shrouds inside the JSC Chamber to provide the desired thermal
environment, while Figure 3(b) displays the equivalent thermal model. The outermost wall is the ambient temperature
pressure vessel; the LN2 Shroud is the first layer inside the chamber walls, flowing liquid nitrogen at approximately
77 K during cryo-stable conditions; the innermost layer is the Helium Shroud which cycles gaseous helium at 20 K
through a 12 kW capacity helium refrigerator. During the test, the gaseous helium from the refrigerator that cools the
innermost shroud was distributed to 16 zones through a valve system. The helium shroud used seven zones, and the
remaining nine zones were plumbed to various GSE. As all of the gaseous helium within the shroud originated from
International Conference on Environmental Systems
5
Figure 2. The OTIS CV Payload Thermal Geometry Model, Showing Major Payload Components
Figure 3. OTIS CV Test Configuration (a) as represented in the CAD model, (b) as represented in the
thermal model with chamber and shrouds, (b) denoting components in the thermal model within the helium
shroud
International Conference on Environmental Systems
6
a single refrigerator, the gas inlet temperature delivered to all zones was the same. Some zones additionally had line
heaters to fine-tune the temperatures reaching their respective GSE.
Harris Corporation developed the complex
suite of thermal GSE inside the helium shroud, as
shown in Figure 3(c), to provide the correct OTIS
payload thermal boundaries11. The Space Vehicle
Thermal Simulator (SVTS) simulated the correct
thermal heat loads of the JWST spacecraft and
harness interfaces in the absence of the final flight
spacecraft bus and sunshield. The SVTS
insulation was closed out to the DTA. A truncated
Layer-5 (L5) sunshield simulator attached to this
“hub” region around the DTA and allowed for
flight-like reflections on the payload. Mounted on
frames from the helium shroud floor, a series of
extremely high emissivity radiator surfaces
known as the Deep Space Environment Radiators
(DSERs), provided localized cold sinks for both
the ISIM enclosure and IEC, and allow for
separate radiative control of these components.
Each DSER was a Z307-painted open aluminum
honeycomb plate which measured an emissivity
of 0.98 at ambient conditions and 0.95 at cryogenic temperatures. The IEC DSER was controlled by a separate helium
circuit to provide appropriate sink temperatures to the ambient-temperature IEC and installed electrical boxes during
transitions, and also to provide for emergency safing in the event of specific unplanned failures of GSE and/or flight
systems in test. Finally, the OTIS payload itself was supported on the massive stainless steel Hardpoint and Offload
Support Structure (HOSS) through a series of monopod and bipod hardpoint struts, which interfaced the BSF at the
–V3 and +V3 ends on either V2 side. The HOSS was helium-cooled and suspended from the ceiling of Chamber A
through a series of telescoping rods, which accommodated structural shrinkage at cryogenic temperatures. Other
support equipment, such as the Center of Curvature Optical Assembly (CoCOA), Auto-Collimating Flats (ACFs), and
Photogrammetry Cameras (PGs) did not have a thermal function, but were employed to conduct optical testing of the
OTIS payload. The Aft Optical Subsystem Source Plate Assembly (ASPA), separately developed by BATC, was
mounted on the OTIS payload Aft Optics Subsystem forward bulkhead to provide sources for optical testing. The as-
built configuration of the OTIS test is shown in Figure 4.
In the OTIS CV test, the primary source of heat on the OTIS payload was from the IEC, kept at 278 K to ensure
the instrument electronics boxes remained within their operating temperatures. The –V3 side of the payload also saw
heat from the SVTS and L5 radiating to the payload thermal management system (TMS), which isolates the cold
optics and instruments. The IEC rejected most of its heat in the –V1 direction through its radiator to the IEC DSER.
Any harness parasitic heat loads conducting from the IEC were further rejected via the Harness Radiator to its own
DSER before they entered the cryogenic instruments. Within ISIM, the MIRI optical bench was actively cooled to
6 K via a cryo-line leading to a GSE cryocooler. The Near IR instruments and the MIRI instrument kinematic mounts
were passively cooled via high-purity Aluminum 1100 heat straps to their respective radiators on the FIR and ADIR.
The ISIM DSERs facing these instrument radiators provided the final sink for payload heat. In flight, the FIR and
ADIR provide ample cooling for the ISIM instruments to reach their operational temperatures, but the cool-down
duration, at approximately 90 days, was not acceptable for ground testing. To accelerate the temperature transition to
instrument operational temperatures, additional ISIM GSE “pre-cool” heat straps were attached to the flight heat-strap
radiator interfaces, allowing for added conductive heat flow out to a GSE helium line, which is especially significant
at low temperatures where radiation heat transfer is less effective. These pre-cool straps were made adiabatic (“zero-
Q’d” i.e., thermally non-conducting) by heaters when the payload approached and achieved thermal balance
conditions by matching the temperature profile in the pre-cool straps to the temperature of the flight hardware. The
hardpoint struts which supported the BSF had similar zero-Q heaters to prevent non-flight-like heat transfer across the
GSE-payload interface. The primary heat paths through the OTIS test are shown in Figure 5.
Figure 4. The test configuration for the OTIS Payload inside
NASA Johnson Space Center's Chamber A Cryogenic
Vacuum Facility
International Conference on Environmental Systems
7
Figure 5. Map of Primary Heat Flows within the OTIS CV Test Configuration
III. OTIS CV As-Run Temperature Profile
The OTIS CV Test was budgeted for 93 days from the pre-test chamber pumpdown to post-test chamber re-
pressurization. A series of 92 thermally-related hardware L&Cs governed the cooldown, cryo-stable, and warmup
phases of the test. The goal of the pre-test model was to develop an optimized method to control the test environment
which transitioned the payload as quickly as possible from ambient conditions to cryogenic operating temperatures,
then back to ambient, while preventing any L&C violations. While certain test phases could be optimized, others
contained so many competing constraints that there was only one way to execute the thermal transition while ensuring
hardware safety. As a testament to the success of the OTIS CV test model, the as-run test transient schedule nearly
exactly matched the pre-test predictions.
After chamber door closure, the pressure inside the chamber dropped to the 10-6 torr scale, after which ambient
vacuum testing of the instruments was performed. The scavenger plates inside the chamber were lowered to LN2
temperatures to capture any molecular contamination. However, the cold sink provided by the scavenger plates in the
chamber resulted in some temperature and gradient violations on the payload. Initially, it was not thought that there
would be a significant ambient impact on the instrument L&Cs due to the small size of the scavengers relative to the
OTIS payload. As such, this case was not part of the transient thermal simulation. During the multiple days of ambient
testing, though, some components cooled enough to enter a range where molecular contamination constraints became
a concern. Entering this range was not intended in the test plan until the beginning of the cooldown phase, where the
temperature transition of the helium shroud environment and DSERs would allow the contamination constraints to be
maintained by the natural gradient formed in the payload. Thus, at ambient, prevention of contamination of instrument
optics necessitated instrument heaters to be turned on to prevent the optics from becoming colder than their
surrounding environment.
Cooldown of the helium shroud began on day 8 of the OTIS CV test. The OTIS helium shroud was in-practice
limited to a rate of 1.5 K/hr, following a safety standard set by the Thermal Pathfinder test. Pre-test predictions of the
transition with the entire OTIS flight payload, however, showed that for the first 80 hours of cooldown in OTIS, the
helium shroud rate of cooling should be decreased to 1 K/hr due to pre-test predictions of the intra-component
International Conference on Environmental Systems
8
gradients on the PMSAs. Below 210 K on the helium shroud, it was predicted that the PMSA structural L&Cs would
have peaked and begun decreasing. Indeed, this trend was observed during test, and past day 3, the helium shroud
could transition to its faster cooldown rate. By day 16, the helium shroud reached 20 K, establishing the planned
environment for cryo-stable testing. Due to the large thermal mass of the OTIS payload, complete cooldown of the
flight hardware required an additional three weeks to reach expected stability at cryogenic temperatures (also referred
to as “cryo-stable temperatures”). As seen from Figure 6, many of the more massive components, such as the primary
mirrors, rapidly cooled at the outset but took up to 32 days from the start of cooldown to begin to exhibit asymptotic
behavior to their steady cryogenic temperatures and achieve the 27 mK/hr stability required for optical testing. The
lower thermal mass Near-IR science instruments (NIRCam, NIRSpec, and FGS) reached a planned contamination
avoidance hold between days 15 and 20 of the total profile. As water contamination was emitted by the composite
ISIM structure between 140 K and 170 K, the instrument optics were held above this temperature range to prevent
contamination within the ISIM enclosure. Conversely, the MIRI bench, with its extreme thermal isolation, was slowest
to cool during this period. Once the ISIM composite structure elements reached below 140 K and no longer emitted
water, the MIRI cryocooler could be turned on to actively drive the MIRI bench temperature to 170 K. When MIRI
cooled to the level of the Near IR instrument temperatures at the threshold of the water contamination band, all SI
bench heaters were stepped down in 5 K increments to prevent cross-contamination from instrument-to-instrument
until they were all below 140 K. The FSM, a mirror having a direct view to all of the instrument pick-off mirrors
(POMs), was also subject to the same contamination requirements as the instruments themselves. The FSM was
similarly held at the median temperature of the SI benches, then stepped down until water deposition was no longer a
concern.
Figure 6. As-Tested Data: OTIS CV Temperature Profiles for Major Components
International Conference on Environmental Systems
9
Past day 20 of the OTIS test, all instrument decontamination heaters except the NIRSpec FPA were turned off to
allow cooling of the instruments via their respective radiators to their cryogenic operating temperatures. At the heat
strap-radiator interface, the “pre-cool” GSE heat straps provided additional conductive cooling to a separate helium
line. The NIRSpec FPA heater remained on above the OA until day 32 to prevent any excess structural gradients,
rapid temperature rate changes, and contamination between both assemblies, as the FPA’s thermal mass was much
lower than the OA. When the instruments reached their operational temperatures between 35 K and 42 K, the zero-Q
heaters for each pre-cool strap interface were turned on. The MIRI cryocooler drove the MIRI bench to about 18K,
after which the cryocooler transitioned from an 18 K Pulse tube cooling method to a 6 K Joule-Thomson process, as
shown in the rapid drop of MIRI bench temperatures around day 34. Once operational temperatures and stability
requirements were achieved on all of the primary optics at Day 40, the cooldown was considered complete and cryo-
stable optical testing could begin.
Day 40 through 72 of the OTIS CV test saw an assortment of instrument and optical tests on the OTIS payload to
understand the telescope’s end-to-end optical performance. Thermal distortion tests were also executed at the end-of-
cooldown to understand the figure drift with temperature change. Oscillations in temperature, as seen most
significantly between days 28 and 50 on the SMA, were caused by the exposure of the payload to the room-temperature
CoCOA aperture for interferometry measurements during optical and thermal distortion testing. From days 37 to 41,
Hurricane Harvey hit the Houston, TX area, causing significant flood and wind damage12. Thankfully, NASA JSC did
not suffer a power outage and the OTIS CV test was impacted only with a five-day optical test schedule delay. Aside
from the off-nominal weather event, the cryo-stable phase was a relatively quiescent period for the thermal subsystem,
and valuable information could be gained about the thermal performance. The ISIM thermal balances were performed
in this phase during periods of minimal instrument activity, and permitted accurate measurements of flight heat strap
conductance and radiator heat dissipation. From Day 72 through 76, the payload underwent a thermal balance of the
OTE subsystem to understand the system-level cryogenic performance of the OTIS payload and provide data to
selectively correlate the OTIS CV thermal model.
At Day 76 of the OTIS CV test, the cryogenic test portion was determined complete and the instruments were
actively heated to 50 K in preparation for payload warmup to ambient. As the helium shroud was slowly increased in
temperature, the sublimation of nitrogen off the payload and helium shroud onto cold scavenger panels between the
helium and LN2 shrouds caused a small and expected pressure rise inside the chamber. As the helium shroud
temperature continued to increase at 1.5 K/hr, the shroud average peaked at 95 K before decreasing at 1.5 K/hr again
to a 75 K plateau. This was in the interest of overdriving the payload for the thermal distortion alignment drift test,
which required measurements of structural temperature gradients and spatial deformation at two separate, stable states:
cryo-stable temperatures and 75 K average. The NIRCam, NIRSpec OA, FGS, and MIRI benches were consistently
held at 10 K above the temperature of their respective DSERs, whereas the instrument pick-off mirrors were 18 K
above the DSERs, and the NIRSpec FPA was 22K above them. Contamination requirements were conflicting: the
DSERs could not be at risk of contaminating the ISIM enclosure by being warmer than the instruments themselves,
nor could any of the instruments risk contaminating others by departing in temperature from the bulk average of the
others. Therefore, through the entirety of warmup, all of the SIs and their optics, as well as the FSM which has a direct
line of sight to the SI POMs, required active control to keep within 15 K of each other.
Following the end of the thermal distortion alignment drift test, day 81 saw the helium shroud resume its steady
1.5 K/hr increase towards 140 K. A short 24-hour hold at 120 K allowed for gradients within the NIRSpec bench to
isothermalize. During this time, a series of latch and hinge deployment tests on the DTA, PMBSS, ADIR, and SMSS
structures raised their respective deployment heaters to 300 K to ensure functionality during flight deployments of the
JWST observatory. At 140 K, a longer helium shroud hold was executed as the more massive optical components
caught up to the helium shroud temperature. The Cryo-Pumping Panels (CPP), cold contamination collectors between
the LN2 and helium shrouds, were raised in temperature at a controlled rate during this time to sublimate their collected
residual air-leak N2 and allow them to be pumped out via the chamber cryo-pumps. An expected rarefied “burp” of
nitrogen gas was released inside the chamber, and in general the increased gas coupling between surfaces lowered
payload gradients. However, this event also caused unexpected approaches towards temperature limits as greater
coupling between surfaces caused increased heat transfer between the composite shell of the room-temperature IEC
and its cold environment. Mitigation steps were taken to halt the CPP warm-up and allow the pumps to lower the
pressure. When the pressure was at an acceptable value and the risk of hardware damage was minimized, the CPPs
continued to warm until the completion of the “burp.” This mitigation prevented the IEC shell temperature from
International Conference on Environmental Systems
10
exceeding its limit. Afterwards, the helium shroud resumed its temperature increase, albeit at a slower rate now due
to contamination requirements. Between 140 K and 170 K, water release from the composite structure could act as a
contaminant for the optical surfaces, and therefore the PMs, SM and TM could not lag the helium shroud by more
than 35 K to reduce any chance of deposition in this contamination “band.” Above 220 K, the release of molecular
contaminants further posed a risk to the optics. Hence, for the entire temperature regime from 140 K to ambient, the
helium shroud rose at the rate of 0.6 K/hr to 0.8 K/hr to ensure that all optical components and instruments would be
able to consistently maintain their required ΔTs with respect to the helium shroud. When the SIs finally reached above
295 K, their heater powers were cascaded down to zero to meet the helium shroud temperature. The helium shroud
was slightly overdriven to bring all of the lagging optical components to above 290 K, the requirement for the end of
the warmup phase. From day 96 onwards, the post-vacuum ambient functional testing phase was executed for an
additional four days before declaring the end of the test.
IV. Thermal Model Performance against Test Data for Cooldown and Warmup
The comparisons for major components between pre-test temperature predictions and the measurements taken
during the OTIS CV test are shown in Figure 7 for the ISIM and Figure 8 for the OTE. Dashed lines represent pre-test
predictions and solid lines represent test measurements. The operational assumptions used in the pre-test predictions
were not always consistent with the as-run operations. For instance, thermal related contamination control L&C’s
were modified in real time during the test; other significant deviations are explained below. Despite this, the overall
comparison between major component temperatures during cooldown was very good except for a PMBSS structure
maximum temperature anomaly discussed later in this section. The model captured both transient and steady-state
cryo-stable behavior of the payload well, and was able to anticipate all major areas of concern with respect to L&Cs,
allowing the OTIS team to plan well in advance the correct environmental controls to ensure hardware safety.
Figure 7. Temperature Comparison for ISIM Components between Model Predictions (Dashed) and Test
Measurements (Solid) during Cooldown
International Conference on Environmental Systems
11
The prescribed helium shroud cooldown profile, determined by pre-test predictions, was matched extremely
closely by the JSC facilities team, including the helium shroud rate transition past day 3 of cooldown. The actively-
heated instruments all followed their prescribed cooldown profiles closely, indicating that the thermal mass and heat
paths out of each of these instruments was captured correctly by the thermal model. The near-IR instruments were all
held at above 170 K starting on day 7 for the ISIM decontamination hold, then stepped down on day 12 through 14
after MIRI was actively cooled to be in-family with the near-IR instrument temperatures. The cooldown of MIRI
below 140K is governed specifically by the operation of the cryocooler. MIRI operated in its Pulse-Tube cryocooling
mode with cold helium circulation through the cryocooler line and MIRI bench, bypassing the Joule-Thomson
cryocooler, until the MIRI optical module reached 18K. Below 18K, a “pinch-point” occurred where circulation was
disabled to the bench and the Joule-Thomson cryocooler was turned on, driving the bench to 6K. As the model
predicted pinch-point occurring at day 22, whereas in test it actually occurred on day 26, there is a perceived
discrepancy in time between model and test. However, this is not an inaccuracy in the model predictions: since the
MIRI cryocooler is an actively-controlled thermal system, the transition of modes at pinch-point is purely determined
by the parasitics on the cryocooler line and the judgment of the cryocooler team, rather than a predetermined time.
The FSM, which tracks the instrument cooldown, was actively controlled via a baseplate heater to achieve all
contamination-prescribed ΔTs with respect to the science instruments. However, the heater was not directly mounted
to the FSM mirror substrate that it sought to control and the weak thermal coupling between the FSM baseplate and
the substrate made it difficult to center the FSM substrate temperature to the median of all of the instrument
temperatures. Active control of the FSM and instrument heaters were not needed when all of these components were
below 140 K, and indeed their cooldown profiles were better matched by the thermal model after day 14.
Figure 8. Temperature Comparison for OTE Components between Model Predictions (Dashed) and Test
Measurements (Solid) during Cooldown
For the OTE optics, the primary mirror predictions followed very closely from ambient temperatures through 170
K. However, below 170 K, especially on the colder PMs, the predictions began diverging significantly from test
measurements, up to 15 K difference at 90 K average temperature. Towards the cryo-stable phase, the predictions and
measurements converged again. An investigation into the difference between model predictions and test in the
temperature regime between 60 K and 170 K pinned the model error to temperature-dependent emissivity assumptions.
International Conference on Environmental Systems
12
This was recognized and studied earlier in the model development and in the interest of runtime for the OTIS CV
thermal model, only two emissivity sets were deemed sufficient: a room-temperature emissivity set when the PMBSS
average was above 90 K, and a cryo-emissivity set below 90 K. Separate radiation conductors were calculated for
each set, and the model transitioned between these sets of conductors once it determined that the PMBSS average fell
below 90 K on cooldown. However, this simplification failed to account for the complex interactions between those
surfaces with temperature-dependent emissivities, especially in this temperature transition regime. It also did not
consider that, at any given time window in the transition, the different cooldown rates of each surface resulted in
higher temperatures and emissivities for more thermally massive components, and lower temperatures and emissivities
for lighter hardware such as soft structure. Therefore, the assumption of room-temperature emissivities on all surfaces
was entirely valid only when all components were at room temperature, and the same statement for cryogenic
emissivities could only be made when all surfaces are at cryogenic steady-state. When the model predicted transient
temperatures on the payload before the emissivity set transition, it overestimated the emissivity on all components,
but especially those components which were less massive and reach colder temperatures quicker. Conversely, after
the emissivity set transition, the model underestimated the emissivity on all components, but especially those
components which were more massive and still warm even after the emissivity switch at 90 K PMBSS average. Hence,
this explained the trend observed on the PM substrate minimum temperature between 60 K and 170 K: before the
emissivity transition, the model overestimated the PM emissivity and predicted colder than test measurements; after
the emissivity transition, the model underestimated PM emissivity, resulting in a slower cooling rate than measured
and allowing for model and test results to converge towards cryogenic steady-state. A similar trend was seen on the
secondary mirror substrate, but this divergence between model and test was both due to the emissivity transition as
well as another factor: the use of the secondary mirror heater. Initially in the OTIS CV test, the SM had a stringent
constraint for allowable temperature differences between its constituent components, each with different thermal
masses. The emissivity simplifications reduced the model’s accuracy to predict cooldown rates for each individual
SM component, and did not detect that there would be an exceedance of SM component-to-component L&Cs in this
transition regime. Since in the test it was observed that ΔTs between SM components were much larger than predicted,
it became necessary to power on the SM heater to drive down intra-SMA gradients. While later it was found that the
incorrect L&C was being applied to the SM in test and as such it did not require active heater control, the prediction
inaccuracies due to emissivity were still a significant contributor to the large divide between model and test
temperatures, in addition to the positive control exerted by the SM heater.
The correct prediction of ΔT between maximum and minimum temperatures on the PMBSS structure was critical
for the OTIS CV test since there was a very stringent temperature-dependent allowable ΔT requirement set forth from
earlier subsystem-level testing, and any exceedances of this allowable limit place the composite structure at risk of
hardware damage. The pre-test predictions for the cold end of the PMBSS structure generally matched well with test
measurements. Certain less-massive components, such as those composite beams and structural extensions with a
direct view to the cold helium shroud, reflected the same error between model and test as the PMs due to the emissivity
set simplification, but were less severe in temperature difference than the PMs. However, on the PMBSS structure
maximum temperatures, the model predictions started diverging versus test at about 120 K, and reached a measured
quasi-steady value at cryo about 22K higher than model predictions. This egregious difference between model and
test results was extremely surprising to the OTIS team at first, given that the model generally predicted within ±3K
for all other major components. Subsequent investigation found that this temperature difference was restricted to the
launch release mechanism (LRM) interface between the room-temperature IEC and the cold BSF, at a titanium
mechanical flexure. The bulk of the other sensors spread through the PMBSS structure showed close matches between
model and test, and assessments by the structural team proved that this localized hot spot at the LRM interface had a
negligible impact on the structural integrity of OTIS. Additional investigations found that about 7-8 K of the
discrepancy between model and test could be attributed to an error in modeled conductance through this interface.
Due to this area being sandwiched between the hot electronics boxes and cold BSF structure, a 160K gradient existed
in this region, and small changes in model conductance values resulted in large changes in predicted temperature
values. However, an additional 11 K of discrepancy is still unaccounted for even with currently-implemented thermal
model fixes; the investigation is still ongoing as to determining the root cause of conflict between model and test at
this interface.
The overall accuracy of the model at thermal balance in cryogenic temperatures versus test measurements is shown
in Figure 9. In total there were over 1000 payload thermal sensors on this very large test article. This histogram shows
the ΔT of model predictions minus test data for all hard structure sensor locations. The steady-state comparison of
International Conference on Environmental Systems
13
model to test showed extremely good correlation: over 90% of
all test sensor measurements were matched within ±4 K by
model-predicted values at corresponding sensor locations. This
gave strong confidence that the OTIS CV thermal model
captured accurately the payload behavior in the cryo-stable test
environment, and further provided assurance that the flight
model would be an accurate predictor of temperatures on-orbit,
since the OTIS payload model was directly sourced from the
flight model. The bulk average of all of the predicted
temperatures to their corresponding measured sensors was
extremely close: within 0.1 K. Even so, it was noticed that
sensors around the SVTS, IEC, and DTA region predicted
slightly warmer in the model than in test. This effect was
eventually attributed to the blanket performance on the IEC. The
OTIS payload model assumed a conservative ε* value of 0.03
for IEC MLI through-thickness effective emissivity, carried
over from the flight observatory model. This allowed more heat
to escape from the warmest payload components inside the
chamber (the room-temperature instrument electronics boxes)
and raised the background sink temperature of the entire
modeled test environment. However, in test, the IEC released
less heat than expected to the environment due to two factors:
the conductance from the IEC to the supporting GSE struts was
correlated from test data to be greater than predicted, reducing
the amount of radiated heat; and the assumed ε* value of 0.03
was too large given the BOL properties of the blankets. This
contributed to a lower background sink temperature. Hence, the
sensor measurements over the payload were consistently lower
in temperature than model predictions.
On warmup, the response of the payload was less well-matched temporally between the model and test results, as
shown in Figure 10 for the ISIM and Figure 11 for the OTE. In the first phase of warmup, the thermal distortion
alignment drift test was executed as the PMBSS and SMSS were driven as quickly as possible to sub-100 mK/hr
stability at a 75 K average, comprising a “State 2” stable condition for photogrammetry measurements of the OTIS
payload structural geometry. The model predicted that this overdrive maneuver on the structure would necessitate the
helium shroud increasing at 1.5 K/hr to a 105 K average, then decreasing to 75 K, after which it would hold constant
at 75 K until both the PMBSS and SMSS were below their temperature rate requirements for “State 2” data acquisition.
In actuality, the payload structure responded much more quickly than expected. From pre-test analysis, it was
understood that a “trigger point” for reversing the helium shroud rate occurred when the PMBSS structure average
reached 71 K, corresponding to the time when the shroud peaked at 105 K. In test, the PMBSS reached this average
temperature when the helium shroud was only at 95 K; this required immediate response from the thermal team to
turn the helium shroud rate to -1.5 K/hr and stem the rise of the PMBSS and SMSS temperature, as well as to prevent
a much longer recovery time to stability. Furthermore, during the hold at 75K, the payload reached its thermal stability
much faster than expected. The entire alignment drift phase completed one full day earlier than planned: it was found
that the faster overall payload response and inability of the model to predict this schedule accurately was again due to
the cryogenic-to-room temperature emissivity set transition. As the payload rose above cryo-stable temperatures, the
assumption of cryogenic emissivities was no longer completely valid, and therefore the model underestimated the
amount of heat transfer actually occurring on the payload. As such, the model response was slower and more
temporally conservative than what was observed during the test.
Beyond the end of the thermal distortion phase on day 5, the test results matched model predictions fairly closely.
The duration of the helium shroud rise to 140 K and subsequent payload response was similar to expectations from
pre-test analysis, save some differences in localized PMBSS temperature rise due to deployment heater testing. The
length of the 140 K helium shroud hold, however, differed due to the risk mitigation steps taken as a result of increased
coupling of payload components during the large N2 “burp”, as discussed in the previous section. The model did not
Figure 9. Comparison of Model Predictions and
Test Data at Thermal Balance
International Conference on Environmental Systems
14
simulate the complexity of free-molecular then convective heat transfer during this event and therefore did not capture
the correct temperature distribution across the payload before the start of the next phase. Into the water contamination
emission band between 140 K and 170 K and during the molecular contamination emission band above 220 K, the
model consistently predicted a slower warmup rate than was executed during the test. The helium shroud warmup rate
was driven by the contamination constraint to maintain all of the OTE optics within 35 K of the average helium shroud
temperature during this phase. The original pre-test thermal analysis stacked worst-case conditions to predict warmup
through these contamination bands with ample schedule conservatism, excluding the benefits of gas heat transfer
during the nitrogen burp to bring the mirrors closer to the helium shroud temperature, as well as allowing for interim
holds to keep all of the instruments at least 10 K above their adjacent DSERS temperatures in accordance with
contamination limits. Given these constraints, it was determined through thermal model predictions that the helium
shroud rate would need to decrease to as low as 0.5 K/hr average to maintain the coldest PM no lower than 35 K below
the helium shroud and “ride the yellow line” with respect to contamination limits. When it was observed in test that
all of the OTE optics and instruments could maintain a faster rate with respect to the helium shroud without violating
any constraints, and it was deemed acceptable by the contamination control team that the instruments only lead the
ISIM DSERS by 6.5 K in the interest of schedule, the helium shroud was sustained at 0.6-0.8 K/hr through the end of
warmup. Therefore, the completion of warmup actually occurred almost two days earlier than pre-test predictions.
Figure 10. Temperature Comparison for ISIM Components between Model Predictions (Dashed) and Test
Measurements (Solid) during Warmup
Two OTIS model configuration-specific issues most affected the predictive capability of the thermal model and
increased the complexity and difficulty of executing the test: (1) the FSM and TM imposed heater powers, and (2) the
variations between PMSA’s modeling detail. Regarding the first issue, the FSM and TM assemblies only contained
one sensor each, a decision made early in the JWST project due to the scarcity of telemetry channels. However, there
were also extremely stringent L&Cs imposed on both components, both for structural and contamination reasons, that
arrived late in manufacture and which required knowledge of temperatures on various subcomponents of each mirror
assembly. As there was no sensor on either mirror substrate, a real-time, reduced thermal model13 was created for both
the FSM and TM which, when operated in conjunction with the OTIS CV test data acquisition system and temperature
telemetry from sensors surrounding each assembly, provided an estimate to the mirror substrate temperatures with a
International Conference on Environmental Systems
15
temperature-varying uncertainty band every two minutes based on the past couple of timesteps of data acquisition.
Temperature deltas both between mirror assembly components and from mirror substrates to the helium shroud could
be calculated with these inferred “virtual sensor” temperatures.
Figure 11. Temperature Comparison for OTE Components between Model Predictions (Dashed) and Test
Measurements (Solid) during Warmup
Both the FSM and TM substrates had the added disadvantage of indirect control with thermal hardware: since
there was a project-level requirement to forbid mounting of heaters directly to the mirror substrates, the FSM heater
was mounted to its mirror assembly baseplate, whereas the TM heater was mounted to its assembly sub-bench. Both
heaters, originally installed for transient test control, could only indirectly affect substrate temperatures through weak
linear and radiative coupling, rather than direct conduction. This contributed to a slow substrate response time to
changes in heater power in both assemblies, and exacerbated the temperature gradients between components when the
heater was turned on. While the “virtual sensors” provided feedback for heater operations, there was also a reliance
on model pre-test predictions to determine the heater power per test phase of the test to ensure that the FSM and TM
complied with all L&Cs. The FSM heater was operated for both cooldown and warmup, while the TM heater was
only used with a low power input during warmup to prevent it from becoming the schedule-driving coldest mirror.
The OTIS CV test, both in cooldown and warmup, saw the FSM requiring significantly more heater power at the
FSM baseplate than model pre-test predictions had prescribed. Especially during test phase transitions, the model
predictions for heater power required (such as when the FSM tries to follow the ISIM bench values during the
decontamination hold in cooldown) were significantly under-predicted versus what was needed by the hardware.
While some of the discrepancy can be attributed to the faster pace of warmup in the OTIS CV test, post-test
investigations showed the bulk of the problem with the pre-test heater power predictions to be twofold: first, the model
only predicted the heat that was required to dissipate on the FSM baseplate, where the heater was mounted, but
inadvertently neglected the significant I2R losses due to the low thermal conductivity, high electrical resistance section
of the flight harness. The operations team also applied the prescribed heat to the entire circuit, which encompassed
the line lengths from the power source outside of the chamber to the payload; this resulted in less-than-prescribed
power at the FSM baseplate. This increased the required heater power at the source to achieve the desired heater power
International Conference on Environmental Systems
16
at the FSM baseplate, since considerable heat was dissipated in the wires. In an aside, the flight model includes all of
the flight harnessing losses. Second, the model assumed that there could be instantaneous jumps to heater power
whenever transitions occurred, but neglected the limitations of the hardware transition rates. These two differences
between the analysis and the as-run contributed to the inaccuracy to predict prescribed FSM heater power, and resulted
in many ad-hoc and immediate changes to the FSM heater power during the test to ensure that it remained within its
L&Cs.
The second issue from model configuration related to the difference in PMSA models used in test. There were
two PMSA models developed for the OTIS program. Firstly, a detailed model had all of the PMSA component
geometries in three dimensions with 32 nodes per mirror segment. This model was specifically developed to address
the modeling accuracy needed in a faster and more stressful transient environment in test than on orbit. The other
model is a more reduced PMSA model, which only had one node per mirror segment and simplified two-dimensional
representations for all of the PMSA components. The simpler representations were the historical models that sought
to reduce run times and focus on cold, steady state conditions. Originally, the motivation for implementation of the
simplified PMSA model stemmed from its usage in flight model predictions, in which commissioning is a longer
duration than test could afford, and so flight steady-state runs are sufficient to capture PMSA behavior with accuracy
and are much faster to run. In the OTIS CV thermal model, which was made more detailed, four segments were
modeled with the detailed PMSA, whereas the rest use the simplified PMSA. It was intended during OTIS test model
development that the detailed model would be required for all 18 segments to capture PMSA behavior during
transitions. However, replacing all segments with detailed PMSA models made model size prohibitive and severely
slowed down runtime due to the ballooning number of radiation conductors from the detailed mirror segments to their
surrounding test environment. It was therefore decided that only the four segments with the most extreme responses
to their surrounding environments (i.e. largest component-to-component ΔTs) would be replaced with the detailed
models as a high-fidelity witness for PMSA intra-component gradients, whereas the rest would remain simplified. At
cryo steady-state, both the simplified and detailed models produced comparable results, and replacing just four PMSAs
with detailed models was a reasonable compromise to make. In transition during cooldown and warmup, though, this
became a much more difficult modeling decision to justify: the difference in PMSA intra-component gradients was
extremely large when predicted by the detailed or simplified models, especially at the beginning of cooldown, when
allowable ΔTs were most restrictive. The detailed PMSA predictions, while not capturing the magnitude of the largest
ΔTs between PMSA components, at least managed to capture the shape of the ΔT curve correctly to what was observed
in test. The simplified model, however, was completely incorrect in both the shape and magnitude of the curves, and
proved useless as a predictive tool for monitoring compliance to the PMSA L&Cs in a highly transient environment.
De-ratings were built into the test design in advance and the magnitude of the PMSA ΔTs never exceeded 60% of
constraint plus margin, and hence there was no hardware safety issue posed by the inaccuracy of the simplified models.
Had the hardware response been larger than observed, this shortcoming in the thermal model may have caused
significant schedule delays in slowing down the helium shroud transition rate to prevent hardware damage.
V. Conclusion
A large thermal modeling effort was performed preceding the JWST OTIS CV test at NASA JSC to predict
schedule and ensure hardware safety. Despite some discrepancies due to modeling assumptions and simplifications in
the interest of runtime and test schedule, the OTIS CV thermal model in total performed exceedingly well at
understanding the payload behavior during transitions from ambient to cryogenic temperatures, as well as in steady-
state predictions at cryogenic thermal balance, especially appreciating the scale of the hardware and the passive deep
cryogenic cooling design. Modeling the test sequence in exquisite detail for both cooldown and warmup gave the
OTIS test team insight to which of the 92 thermal L&Cs would be schedule drivers in each phase of test, as well as
identified the correct “knobs to turn” to mitigate any exceedance of these L&Cs. The OTIS thermal team spent
enormous effort ensuring that the payload was safe both structurally and contamination-wise before the test, and the
thermal model provided a valuable guideline for what to focus on and how best to avoid any problems. It is due to this
extensive and exhaustive modeling campaign that the OTIS CV test was executed smoothly and successfully. As there
is no observatory-level testing planned for JWST and the OTIS CV test represents the final system-level cryogenic
test of the optics and instruments, the excellent match of the OTIS CV thermal model with test results also provides
confidence to the flight team that the thermal models are accurately predicting the temperatures that will be seen in
flight.
International Conference on Environmental Systems
17
The lessons gleaned from this effort may prove valuable to future cryogenic missions in the interest of creating
better predictive tools for hardware performance during test. Some suggestions for improving future system-level
accuracy of test cryogenic thermal models are as follows. These will also be documented in NASA Goddard Space
Flight Center’s Applied Engineering and Technology Directorate Wiki Lessons Learned database.
For transient cooldown and warmup predictions, especially in the transition region between 60 K and 170 K,
use of more temperature-dependent emissivity sets is critical to accurately capture radiative heat transfer
behavior of hardware, especially with respect to L&Cs. If dynamic calculations of temperature-dependent
emissivity per analysis timestep is not an option due to large model size and prohibitive runtimes, multiple
emissivity sets especially through this transition regime are a basic necessity to ensure that absolute
temperature predictions on radiation-dominant components are accurate.
Blanket performance can be a large contributor to the overall steady-state temperatures. If the observed
temperatures differ from test values, it may be due to overly conservative assumptions about blanket
performance.
Transition to cryogenic temperatures and back is inherently stressful for hardware. For large scale cryogenic
missions like JWST, simulation of accurate radiation heat exchange and temperature gradient-induced
stresses due to the coefficient of thermal expansion (CTE) of the constituent materials may demand more
detailed thermal geometric modeling. While 2D-surface simplifications of critical assemblies may be
sufficient for steady-state flight predictions, it does not capture transient behavior well, especially when
tracking compliance with structural limits between mirror components in a schedule-limited test. When
relying on a thermal model to confirm adherence to L&Cs, it is desirable to retain detailed model geometries
although they result in much longer model runtimes and hence model turn-around times, which then needs
to be accommodated in the analysis schedule. Future large scale cryogenic missions should develop a
modeling and analysis plan that trades the analysis speed and geometric fidelity against the program phase
and accuracy goals.
Due to modeling best-practices targeted towards creating a conservative payload model, and focused on a
test start time tied to payload direct cooling but with an extended hold time near ambient in a 20K space, the
transition times determined in analysis between ambient and cryogenic temperatures may tend to be longer
than actually observed in test. Therefore, a good assumption to make is that overall temperature transition
duration will be shorter than the model predictions.
For areas where the model predictions did not match with test results, such as the LRM interface between the
BSF and the IEC as discussed earlier, a full assessment needs to be made to determine if the observed
temperatures or phenomena have a significant effect on flight performance. For this region specifically, the
location of these LRM interface sensors was bound at their test-measured thermal balance temperatures in
the flight model and analyzed in steady-state to understand the impact on flight metrics. It was found that
this level of static, localized gradient does cause a slight degradation to thermal margins, but does not
meaningfully contribute to the amount of stray light on the optics nor the performance of the instruments.
For interfaces with large thermal gradients or temporal thermal change, a greater number of test sensors is
crucial to understanding the heat flows and physical phenomena that occur at each interface in case the trends
observed do not match what was expected from pre-test predictions. As discussed in the paper, much of the
difficulty with diagnosing the anomaly at the BSF interface to the IEC or understanding of performance
against L&Cs on the FSM and TM substrates was due to an inadequate number of test and flight sensors at
those interfaces. Safety and anomaly assessment are the main reasons cryogenic systems need far more
diagnostic sensors than ambient spacecraft and it needs to be recognized that this needs to be a planned
expense. Furthermore, the mirror heaters being placed on separate components removed from the FSM and
TM substrate made for extremely difficult control of the mirror substrate temperatures. Future missions
should always make considerations for placing heaters on directly on the components they attempt to control,
especially if that component has stringent structural and/or contamination requirements. This will increase
hardware safety, reduce test time that can be paced by analytical uncertainty, reduce reliance and effort from
system level models, and avoid undue effort post-test to troubleshoot anomalies and understand behavior.
Acknowledgments
The authors would like to thank the entire OTIS CV test team for their countless hours of hard work and sacrifice
to ensure a successful test at NASA JSC, especially in the face of extraordinary events such as Hurricane Harvey. The
International Conference on Environmental Systems
18
authors would also like to recognize the following OTIS team members for their invaluable contributions to the
thermal subsystem: Tamara O’Connell and Carl Reis at ATA Aerospace; Rusty Schweickart, Denny Teusch, and
Pamela Brinckerhoff at Ball Aerospace; William Chang, Angelique Davis, Theodore Michalek, and Chuen Or at Edge
Space Systems; Christopher May and Paul Cleveland at Energy Solutions International L.L.C.; Danilo Tagle at
Genesis Engineering Solutions, Inc.; Dwight Cooke, Keith Havey, Jesse Huguet, Clint Travis, and Mark Waldman at
Harris Corporation; the NASA Johnson Space Center Chamber A Facilities Team; Joshua Adamson, George Harpole,
Perry Knollenberg, Peter Munther, and John Pohner at Northrop Grumman Corporation; Sang Park at the Smithsonian
Astrophysical Observatory; Dharmendra Patel at Stinger Ghaffarian Technologies Inc.; and Christine Cottingham,
Russell Harrison, John Johnston, Mark Kobel, Juli Lander, Doug McGuffey, Al Sherman, Regis Venti, and Mark
Voyton at NASA Goddard Space Flight Center.
References 1Feinberg, L., Voyton, M., Lander, J., Keski-Kuha, R., and Matthews, G. “James Webb Space Telescope Optical
Telescope Element/Integrated Science Instrument Module (OTIS) Status.” Proceedings of the SPIE. Volume 9904,
id. 990407. 2016. 2Kimble, R., Apollo, P., Feinberg, L., Glazer, S., Hanley, J., Keski-Kuha, R., Kirk, J., Knight, S., Lambros, S.,
Lander, J., McGuffey, D., Mehalick, K., Ohl, R., Ousley, W., Reis, C., Reynolds, P., Begoña Vila, M.; Voyton, M.,
Waldman, M., Whitman, T. “Cryo-Vacuum Testing of JWST’s Integrated Telescope & Scientific Instrument Suite.”
American Astronomical Society. AAS Meeting 229, id. 238.14. 2016. 3Park, S., Brinckerhoff, P., Franck, R., Schweickart, R., Thomson, S., Burt, W., and Ousley, W. “Successful
Completion of the JWST OGSE2 Cryogenic Test at JSC Chamber-A while Managing Numerous Challenges.” 46th
International Conference on Environmental Systems. Vienna, Austria, July 13-17, 2016. 4Ousley, W., Burt, W., and Davis, A. “James Webb Space Telescope Cryogenic Thermal Pathfinder Test” 44th
International Conference on Environmental Systems. Tucson, AZ, July 13-17, 2014. 5Davis, A., Ousley, W., Burt, W., and Cottingham, C. “James Webb Space Telescope Thermal Pathfinder Test
Development.” 46th International Conference on Environmental Systems. Vienna, Austria, July 13-17, 2016. 6Davis, A., Ousley, W., and Burt, W. “James Webb Space Telescope Thermal Pathfinder Test Results.” 47th
International Conference on Environmental Systems. Charleston, SC, July 17-20, 2017. 7Glazer, S., and Comber, B. “James Webb Space Telescope Integrated Science Instrument Module Thermal
Vacuum/Thermal Balance Test Campaign at NASA's Goddard Space Flight Center.” 46th International Conference
on Environmental Systems. Vienna, Austria, July 13-17, 2016. 8Yang, K., Glazer, S., Ousley, W., and Burt, W. “Thermal Considerations for Reducing the Cooldown and Warmup
Duration of the James Webb Space Telescope OTIS Cryo-Vacuum Test.” 47th International Conference on
Environmental Systems. Charleston, SC, July 17-20, 2017. 9Franck, R., Gurule, A., Brown, D., Brinckerhoff, P., McCallan, T., and Renbarger, M. “High Performance
Cryogenic Radiators for James Webb Space Telescope.” 46th International Conference on Environmental Systems.
Vienna, Austria, July 13-17, 2016. 10Homan, J. Cerimele, M., and Montz, M. “Creating the Deep Space Environment for Testing the James Webb
Space Telescope at NASA Johnson Space Center's Chamber A.” 43rd International Conference on Environmental
Systems. Vail, CO, July 14-18, 2013. 11Havey, K., Cooke, D., Huguet, J., and Day, R. “Thermal Management of JWST Cryo Test Support
Equipment.” 48th International Conference on Environmental Systems. Albuquerque, NM, July 8-12, 2018
(submitted for publication). 12Glazer, S., Yang, K., Comber, B., Ousley, W., and Cleveland, P. “Off-Nominal Planning for the Cryogenic
Thermal Vacuum/Thermal Balance Test of the James Webb Space Telescope Optical Telescope Element/Integrated
Science Instrument Module at the Johnson Space Center.” 48th International Conference on Environmental Systems.
Albuquerque, NM, July 8-12, 2018 (submitted for publication). 13Franck, R., Schweickart, R., and Comber, B. “The Use of Real Time Models to Produce Virtual Sensor Telemetry
During the JWST OTIS Test.” 48th International Conference on Environmental Systems. Albuquerque, NM, July 8-
12, 2018 (submitted for publication).